Syed Talal Wasim

About Me

I am a PhD student in the Computer Vision Group at the University of Bonn, supervised by Professor Dr. Jürgen Gall. My research is on video understanding that looks forward rather than backward: anticipating what happens next in a video and, increasingly, generating it. Current work centres on reinforcement-learning post-training for generative video and world models, and on how to give a large multimodal model a new capability from little data without losing what it already knows.

Before the PhD I was an Associate Researcher at the Intelligent Visual Analytics Lab (IVAL) at MBZUAI, working with Dr. Salman Khan on efficient video recognition, adapting vision-language models to video, and open-vocabulary spatio-temporal grounding. Several of those models have since been adopted as backbones and baselines in subsequent video research.

I did my master’s in Image Processing and Computer Vision (IPCV) on an Erasmus Mundus scholarship, with my thesis at EPFL’s CVLAB under Dr. Mathieu Salzmann and an internship at the Empathic Computing Lab under Dr. Mark Billinghurst. My undergraduate degree is in Electrical Engineering from Habib University in Karachi, Pakistan.

I’m glad to hear from prospective students and collaborators interested in any of the topics below.

Research Interests

My work follows one line: from perceiving what is in a video, to anticipating what happens next, to generating what could happen. Three things carry through every step: multimodal alignment, computational efficiency, and adapting foundation models without erasing what they already know.

Anticipation and world models: long-horizon and open-vocabulary action anticipation; generative video and world models; unified models that both understand and generate
Post-training foundation models: reinforcement-learning post-training (GRPO) as targeted property transfer; preserving prior capabilities; self-rewarding formulations that rule out reward hacking
Spatio-temporal grounding and segmentation: open-vocabulary video grounding; memory-efficient long-video object segmentation; referring localization in vision-language models
Efficient video architectures: state-space models for images and video; encoder-free video-language alignment
Multimodal adaptation: prompt learning and regularized adaptation of vision-language models for zero-shot and open-vocabulary generalization

News

Jul 2026
Our paper “Open-Vocabulary Long-term Action Anticipation” is accepted in ECCV 2026.
Mar 2026
Our paper “RedSage: A Cybersecurity Generalist LLM” is accepted in ICLR 2026.
Jul 2025
Our paper “MixANT: Observation-dependent Memory Propagation for Stochastic Dense Action Anticipation” is accepted in ICCV 2025.
Feb 2025
Three of our papers (Video-Panda, GroupMamba, and STING-BEE) have been accepted in CVPR 2025.
Dec 2024
Our paper “Efficient Video Object Segmentation via Modulated Cross-Attention Memory” is accepted in WACV 2025.
Oct 2024
New preprint released titled “Distillation-free Scaling of Large SSMs for Images and Videos”.
Mar 2024
Our paper “VideoGrounding-DINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding” is accepted in CVPR 2024.
Feb 2024
My student Muhammad Zain Yousuf’s bachelor thesis titled “AR-VPT: Simple Auto-Regressive Prompts for Adapting Frozen ViTs to Videos” is accepted in VISAPP 2024.
Jan 2024
I started a PhD at the University of Bonn, Germany working on Long-Term Multimodal Video Understanding, under the supervision of Professor Dr. Juergen Gall.
Oct 2023
Our paper “Hardware Resilience Properties of Text-Guided Image Classifiers” is accepted in NeurIPS 2023.
Aug 2023
Our paper “Video-FocalNets: Spatio-Temporal Focal Modulation for Video Action Recognition” is accepted in ICCV 2023.
Aug 2023
Our paper “Self-regulating Prompts: Foundational Model Adaptation without Forgetting” is accepted in ICCV 2023.
Jun 2023
Our paper “Toward Automatic Typography Analysis: Serif Classification and Font Similarities” is accepted in the Journal of Data Mining in Digital Humanities (JDMDH).
Mar 2023
Our paper “Vita-CLIP: Video and text adaptive CLIP via Multimodal Prompting” is accepted in CVPR 2023.
Jun 2022
Our paper “Using Facial Micro-Expressions in Combination With EEG and Physiological Signals for Emotion Recognition” is accepted in Frontiers in Psychology.
Apr 2022
I started working as a researcher at MBZUAI. I was supervised by Dr. Salman Khan, working on multimodal video understanding.
Jul 2021
I was accepted in the ETH Robotics Summer School and Symposium.
Jun 2021
I defended my master’s thesis and graduated from the IPCV master’s program.
May 2021
Our paper on synthetic data for object detection is accepted to CVPR 2021 CV4Animals workshop.
Feb 2021
I started my master’s thesis in the CVLAB at EPFL supervised by Dr. Mathieu Salzmann. I worked on automated typography analysis on figurative content.
Jul 2020
I started a remote research internship at the Empathic Computing Lab supervised by Dr. Mark Billinghurst.
Sep 2019
I started my master’s degree in Image Processing and Computer Vision (IPCV) funded by the Erasmus Mundus Joint Master’s Degree (EMJMD) scholarship program.
Jun 2019
I completed my undergraduate degree in Electrical Engineering with a Minor in computer science. Graduated first in class with the Dean’s Medal.

Publications

2026

Open-Vocabulary Long-term Action Anticipation

Syed Talal Wasim, Jinhui Yi, Hamid Suleman, Ahmad Javed, Yanan Luo, Muzammal Naseer and Juergen Gall

ECCV, 2026

Paper Project page

@inproceedings{wasim2026openVocabAnt,
  title={Open-Vocabulary Long-term Action Anticipation},
  author={Syed Talal Wasim and Jinhui Yi and Hamid Suleman and Ahmad Javed and Yanan Luo and Muzammal Naseer and Juergen Gall},
  booktitle={ECCV},
  year={2026}
}

RedSage: A Cybersecurity Generalist LLM

Naufal Suryanto, Muzammal Naseer, Pengfei Li, Syed Talal Wasim, Jinhui Yi, Juergen Gall, Paolo Ceravolo, Ernesto Damiani

ICLR, 2026

Paper Project page

@inproceedings{suryanto2026redsage,
  title={RedSage: A Cybersecurity Generalist LLM},
  author={Naufal Suryanto and Muzammal Naseer and Pengfei Li and Syed Talal Wasim and Jinhui Yi and Juergen Gall and Paolo Ceravolo and Ernesto Damiani},
  booktitle={ICLR},
  year={2026}
}

2025

MixANT: Observation-dependent Memory Propagation for Stochastic Dense Action Anticipation

Syed Talal Wasim, Hamid Suleman, Olga Zatsarynna, Muzammal Naseer and Juergen Gall

ICCV, 2025

Paper Project page

@inproceedings{wasim2025mixant,
  title={MixANT: Observation-dependent Memory Propagation for Stochastic Dense Action Anticipation},
  author={Syed Talal Wasim and Hamid Suleman and Olga Zatsarynna and Muzammal Naseer and Juergen Gall},
  booktitle={ICCV},
  year={2025}
}

Video-Panda: Parameter-efficient Alignment for Encoder-free Video-Language Models

Jinhui Yi*, Syed Talal Wasim*, Yanan Luo*, Muzammal Naseer and Juergen Gall

CVPR, 2025

Paper Project page

@inproceedings{yi2025vpanda,
  title={Video-Panda: Parameter-efficient Alignment for Encoder-free Video-Language Models},
  author={Jinhui Yi* and Syed Talal Wasim* and Yanan Luo* and Muzammal Naseer and Juergen Gall},
  booktitle={CVPR},
  year={2025}
}

GroupMamba: Parameter-Efficient and Accurate Group Visual State Space Model

Abdelrahman Shaker, Syed Talal Wasim, Salman Khan, and Fahad Shahbaz Khan

CVPR, 2025

Paper Code

@inproceedings{shaker2024groupmamba,
  title={GroupMamba: Parameter-Efficient and Accurate Group Visual State Space Model},
  author={Abdelrahman Shaker and Syed Talal Wasim and Salman Khan and Juergen Gall and Fahad Shahbaz Khan},
  booktitle={CVPR},
  year={2025}
}

STING-BEE: Towards Vision-Language Model for Real-World X-ray Baggage Security Inspection

D. Velayudhan, A. Ahmed, M. Alansari, N. Gour, A. Behouch, T. Hassan, Syed Talal Wasim, N. Maalej, M. Naseer, J. Gall, M. Bennamoun, E. Damiani and N. Werghi

CVPR, 2025

Paper Code

@inproceedings{velayudhan2024sting,
  title={STING-BEE: Towards Vision-Language Model for Real-World X-ray Baggage Security Inspection},
  author={Divya Velayudhan and Abdelfatah Ahmed and Mohamad Alansari and Neha Gour and Abderaouf Behouch and Taimur Hassan and Syed Talal Wasim and Nabil Maalej and Muzammal Naseer and Juergen Gall and Mohammed Bennamoun and Ernesto Damiani and Naoufel Werghi},
  booktitle={CVPR},
  year={2025}
}

Efficient Video Object Segmentation via Modulated Cross-Attention Memory

Abdelrahman Shaker, Syed Talal Wasim, Martin Danelljan, Salman Khan, Ming-Hsuan Yang and Fahad Shahbaz Khan

WACV, 2025

Paper Code

@inproceedings{shaker2025mavos,
  title={Efficient Video Object Segmentation via Modulated Cross-Attention Memory},
  author={Abdelrahman Shaker and Syed Talal Wasim and Martin Danelljan and Salman Khan and Ming-Hsuan Yang and Fahad Shahbaz Khan},
  booktitle={WACV},
  year={2025}
}

Distillation-free Scaling of Large State-Space Models for Images and Videos

Hamid Suleman*, Syed Talal Wasim*, Muzammal Naseer and Juergen Gall

IJCV

Paper Code

@article{suleman2026stablemamba,
  title={Distillation-free Scaling of Large State-Space Models for Images and Videos},
  author={Hamid Suleman* and Syed Talal Wasim* and Muzammal Naseer and Juergen Gall},
  journal={IJCV},
  year={2026}
}

2024

Video-GroundingDINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding

Syed Talal Wasim, Muzammal Naseer, Salman Khan, Ming-Hsuan Yang and Fahad Shahbaz Khan

CVPR, 2024

Paper Code

@inproceedings{wasim2024vgdino,
  title={Video-GroundingDINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding},
  author={Syed Talal Wasim and Muzammal Naseer and Salman Khan and Ming-Hsuan Yang and Fahad Shahbaz Khan},
  booktitle={CVPR},
  year={2024}
}

AR-VPT: Simple Auto-Regressive Prompts for Adapting Frozen ViTs to Videos

Muhammad Zain Yousuf, Syed Talal Wasim, Syed Nouman Hasany and Muhammad Farhan

VISAPP, 2024

Paper Code

@inproceedings{yousuf2024arvpt,
  title={AR-VPT: Simple Auto-Regressive Prompts for Adapting Frozen ViTs to Videos},
  author={Muhammad Zain Yousuf and Syed Talal Wasim and Syed Nouman Hasany and Muhammad Farhan},
  booktitle={VISAPP},
  year={2024}
}

2023

Hardware Resilience Properties of Text-Guided Image Classifiers

Syed Talal Wasim, Kabila Haile Soboka, Abdulrahman Mahmoud, Salman Khan, David Brooks and Gu-Yeon Wei

NeurIPS, 2023

Paper Code

@inproceedings{wasim2023textres,
  title={Hardware Resilience Properties of Text-Guided Image Classifiers},
  author={Syed Talal Wasim and Kabila Haile Soboka and Abdulrahman Mahmoud and Salman Khan and David Brooks and Gu-Yeon Wei},
  booktitle={NeurIPS},
  year={2023}
}

Video-FocalNets: Spatio-Temporal Focal Modulation for Video Action Recognition

Syed Talal Wasim*, Muhammad Uzair Khattak*, Muzammal Naseer, Salman Khan, Mubarak Shah and Fahad Shahbaz Khan

ICCV, 2023

Paper Code

@inproceedings{wasim2023vfn,
  title={Video-FocalNets: Spatio-Temporal Focal Modulation for Video Action Recognition},
  author={Syed Talal Wasim* and Muhammad Uzair Khattak* and Muzammal Naseer and Salman Khan and Mubarak Shah and Fahad Shahbaz Khan},
  booktitle={ICCV},
  year={2023}
}

Self-regulating Prompts: Foundational Model Adaptation without Forgetting

Muhammad Uzair Khattak*, Syed Talal Wasim*, Muzammal Naseer, Salman Khan, Ming-Hsuan Yang and Fahad Shahbaz Khan

ICCV, 2023

Paper Code

@inproceedings{khattak2023promptsrc,
  title={Self-regulating Prompts: Foundational Model Adaptation without Forgetting},
  author={Muhammad Uzair Khattak* and Syed Talal Wasim* and Muzammal Naseer and Salman Khan and Ming-Hsuan Yang and Fahad Shahbaz Khan},
  booktitle={ICCV},
  year={2023}
}

Vita-CLIP: Video and text adaptive CLIP via Multimodal Prompting

Syed Talal Wasim, Muzammal Naseer, Salman Khan, Fahad Shahbaz Khan and Mubarak Shah

CVPR, 2023

Paper Code

@inproceedings{wasim2023vita,
  title={Vita-CLIP: Video and text adaptive CLIP via Multimodal Prompting},
  author={Syed Talal Wasim and Muzammal Naseer and Salman Khan and Fahad Shahbaz Khan and Mubarak Shah},
  booktitle={CVPR},
  year={2023}
}

Toward Automatic Typography Analysis: Serif Classification and Font Similarities

Syed Talal Wasim, Romain Collaud, Lara Défayes, Nicolas Henchoz, Mathieu Salzmann and Delphine Ribes

Journal of Data Mining in Digital Humanities, 2023

Paper Code

@article{wasim2023gest,
  title={Toward automatic typography analysis: serif classification and font similarities},
  author={Syed Talal Wasim and Romain Collaud and Lara Défayes and Nicolas Henchoz and Mathieu Salzmann and Delphine Ribes},
  journal={Journal of Data Mining in Digital Humanities (JDMDH)},
  year={2023}
}

2022

Using Facial Micro-Expressions in Combination With EEG and Physiological Signals for Emotion Recognition

Nastaran Saffaryazdi, Syed Talal Wasim, Kuldeep Dileep, Alireza Farrokhi Nia, Suranga Nanayakkara, Elizabeth Broadbent and Mark Billinghurst

Frontiers in Psychology, 2022

Paper

@article{wasim2022ecl,
  title={Using facial micro-expressions in combination with EEG and physiological signals for emotion recognition},
  author={Nastaran Saffaryazdi and Syed Talal Wasim and Kuldeep Dileep and Alireza Farrokhi Nia and Suranga Nanayakkara and Elizabeth Broadbent and Mark Billinghurst},
  journal={Frontiers in Psychology},
  year={2022}
}

2021

Sim-to-Real Transfer for Object Detection and Localization on Animals

Syed Talal Wasim, Syed N. Hasany, Kainat Abbasi, Huda Feroz, Anisa A. Ahmed, Mudasir H. Shaikh and Muhammad Farhan

CV4Animals Workshop, CVPR 2021

Poster

@inproceedings{wasim2021cv4animals,
  title={Sim-to-Real Transfer for Object Detection and Localization on Animals},
  author={Syed Talal Wasim and Syed N. Hasany and Kainat Abbasi and Huda Feroz and Anisa A. Ahmed and Mudasir H. Shaikh and Muhammad Farhan},
  booktitle={CV4Animals CVPR Workshop},
  year={2021}
}

Services

Journal reviewer

Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
Transactions on Neural Networks and Learning Systems (TNNLS)
Transactions on Image Processing (TIP)
Transactions on Machine Learning Research (TMLR)
International Journal of Computer Vision (IJCV)
Pattern Recognition

Conference reviewer

Computer vision (CVPR, ICCV, ECCV, WACV, ACCV)
Artificial intelligence and machine learning (NeurIPS, ICLR, ICML, AAAI)

Project supervision

Co-supervise undergraduate projects in computer vision at Habib University
Co-supervise high-school students in Pakistan for the Intel International Science and Engineering Fair (ISEF)